Notwithstanding its title, Yannis Ioannides's densely written, rigorous book is more an analytical survey of recent theoretical and empirical developments in urban economics than a general study of social interactions. Throughout, Ioannides's main goal is to derive and compare the implications of models of rational individual behavior in spatial equilibrium. Social interactions, defined by the author as “direct agent-to-agent interactions that are not mediated by price” (p. xi), play a critical role because they often require or benefit from spatial proximity. From such interactions spring many of the agglomeration effects that give rise to cities and govern their growth and size distribution.

The book's opening chapter starts from first principles, developing models of social interactions and social effects between individuals, and exploring the daunting empirical challenge of distinguishing true social interaction effects, such as the influence of peer behavior, from other sources of correlation between actions, such as common characteristics. Ioannides then proceeds to partial equilibrium models of locational decisions by individuals and firms, including Schelling-type tipping models of neighborhood segregation, and models of agglomeration economies that lead to concentration of firms in the same industry within particular cities or even neighborhoods.

The central chapters (5 and 6) start with the “canonical” Alonso-Mills-Muth model of urban spatial equilibrium, in which residents arrange themselves around a central business district and equilibrium housing rents decline with distance from the city center to offset transportation costs to jobs and amenities. Ioannides goes on to explore a variety of extensions and generalizations of the basic urban model, allowing for positive production externalities across firms that locate near each other, as well as agglomeration economies due to advantages in job search and matching and human capital spillovers. Wages tend to be higher in cities, and especially in cities with a larger stock of human capital, but the difficult empirical question is whether these effects are due to true agglomeration effects and spillovers, or rather perhaps selection effects. Surveying some of the literature Ioannides concludes that “… agglomeration effects are real, even when their elasticities are quite small…” (p. 287).

The final chapters explore empirical regularities in city size and growth and develop models of interurban trade that might account for the number, size, and degree of specialization of cities. Chapter 8 is an extended survey of the work on Zipf's Law of Cities and related empirical regularities, as well as models of city growth that might give rise to Zipf's Law as a limiting distribution. Zipf's Law suggests that the product of a city's population and its population ranking is a constant: thus the second-largest city should be half as large as the largest city, and so forth. Such a power law is an interesting regularity that more or less holds across different definitions of urban agglomerations, countries, and time periods. Explaining why it holds is another matter. When all is said and done, Ioannides quite sensibly writes, “I would argue that one should be suspicious of attempts to characterize city size distributions as obeying immutable laws” (p. 377).

Chapter 9 is the most ambitious in the book, laying out a series of increasingly complex models of intercity trade and urban growth. The author, whether out of sympathy or wit, warns: “Here I am asking the reader to delve into fairly complex material” (p. 410). In his full model, monopolistically competitive firms produce intermediate goods a la the familiar Dixit-Stiglitz model, which are then used to produce two different manufactured goods, in turn used to produce consumption and investment. The model is rich in implications for a “system of cities”: cities may be autarkic, engage in trade of intermediate goods, or specialize in producing one or the other manufactured good and trade that. Cities benefit from growth because larger cities can produce a greater variety of intermediate goods for trade. The evolution of such a system of cities will be shaped, unsurprisingly, by transportation and congestion costs, but the rich range of possible outcomes is a challenge to empirical testing. One quite interesting implication of the model is the parallel between cities and firms. Like competitive firms with U-shaped long-run costs, individual cities grow to the point where congestion costs offset agglomeration effects and constant returns set in; further growth occurs as new replicant cities are born and grow. Thus the model implies constant returns to scale at the level of the national economy, even in the presence of local external economies.

For economic historians, From Neighborhoods to Nations provides an up-to-date survey of a number of theoretical and empirical issues in the study of social interactions and urban economics, but the book is not exactly user-friendly. The proliferation of models is reflected in the book's choppy structure of subsections within subsections. Empirical methods are discussed, but the reader will have to consult the references to find out how to implement them. Perhaps most disappointingly for the historian, Ioannides has relatively little to say about the political economy aspects of cities, such as rent-seeking through town-building or urban machine politics, or city planning and policy. Surely an important aspect of social interaction in cities and the shaping of systems of cities is the political process, including the interaction between levels of government. Policy issues come up here and there, but often in the abstract context of welfare implications of equilibria. Ioannides presents models in which individuals sort into communities within an urban area according to income and taste for amenities (chapter 3), but curiously one finds no mention of Tiebout in either the reference list or the index. Readers interested in such political dimensions of urban economics will have to look elsewhere.